1. Field of the Invention
The present invention relates to an image processing apparatus which combines various predetermined data such as image data, character data and voice data or the like with original image data to generate combined image data and separates the various predetermined data from the combined image data.
2. Description of the Related Art
It has become easy for a large number of people to share the same data through a network such as Local Area Network (abbreviated as “LAN”) or the Internet or the like by digitalization of multimedia data such as an image or a voice or the like.
Contents business draws people's attention as business utilizing a network. Contents business is a business which directly transmits digital data such as an image, a music, a video or the like to a consumer via a network. E-commerce by which commercial products are dealt with on the network makes commerce of the commercial products more efficient, but the e-commerce requires transporting means such as a truck or the like to be used for delivering the commercial products when the commercial products in themselves are corporeal. On the contrary, the contents business does not require the transporting means since the commercial products in themselves are delivered to the consumer via the network. Consequently, pioneering a new business making full use of such feature as mentioned above is expected.
However, digital data can be easily copied and therefore, unless any countermeasure for preventing such a copy as mentioned above is taken, there is possibility that an illegal copy is rampant. An electronic watermark is researched as effective technique for preventing the illegal copy.
The electronic watermark relates to technique for utilizing feature of human perception such as visual perception and auditory perception or the like and embedding different predetermined data from digital contents in themselves such as a still image, a motion image and audio or the like into the digital contents so that it is difficult for a man to perceive the different predetermined data. The electronic watermark cannot directly prevent the illegal copy, but can indirectly prevent the illegal copy by restraining the illegal copy (e.g. as referred to Japanese Unexamined Patent Publication JP-A 2000-106624 (2000)).
To realize the above mentioned electronic watermark, various methods are proposed. For example, it is possible that predetermined data is embedded in low-order bit of a gray scale value of original image data as a simple method of embedding the predetermined data in the original image data.
A sampling point sampled at the time of digitalizing an image is called a pixel. This pixel becomes a minimum unit configuring a digital image. In a monochrome facsimile, each pixel is either white or black and therefore, each pixel has two gray scales and the two gray scales can be represented as one bit by allocating a gray scale values 0 to white and allocating a gray scale value 1 to black. However, with respect to a gray-scale image which is of color, it is required that the gray scale of the pixel is increased. The gray scale value of the color image is often represented as 8 bits in view of easiness of handling the gray scale value on a computer.
By the way, since a cone which is a color sensitive organ in a human's eye consists of three kinds L, M and S (color space for 3 primary colors), three values are required for representing a plurality of colors, but, in other words, existence of the three values is sufficient to represent the plurality of colors. Changing the three gray scale values of red(R), green(G) and blue (B) enables various kinds of colors to be displayed on a display. Therefore, a color image is represented by 8 bit concerning each of RGB.
The low-order about 2 bit of the gray scale value of the color image generally and originally means noise, and therefore, even if the predetermined data is embedded in this low-order bit, there is less deterioration of quality of the color image in the case of small amount of the predetermined data. However, the larger amount of the predetermined data becomes, the greater deterioration of the quality of the color image becomes. However, a method for realizing electronic watermark actually is more sophisticated and retrieving the predetermined data embedded in the original image data cannot be easily performed, and deterioration of quality of the color image is inevitably accompanied. In addition, in advance of compression technique, since the predetermined data is embedded in the original image data and then the whole of the image data is compressed, it is essential that the method for realizing electronic watermark should be an embedding method having good endurance for compression, in other words, a method by which the embedded predetermined information can be retrieved when the whole of the image data is encoded and then decoded.
Since an image has large amount of data, the image is often compressed. A method of compression encoding comprises lossless compression and lossy compression. The lossless compression is conversion by which data decompressed after compression of the data perfectly coincides with data having existed before compression of the data. The lossy compression is conversion by which data decompressed after compression of the data perfectly does not coincide with data having existed before compression of the data and once the compressed data is not perfectly returned. In the lossy compression, a compression ratio is enhanced by leaving information on a part of the display which is sensitively recognized by a human's eye and eliminating information especially on a part of the display which is more insensitively recognized by the human's eye.
The lossy compression uses transform encoding, for instance, with use of discrete cosine transform (abbreviated as “DCT”) and wavelet transform or the like. In the transform encoding, pixels lined on a plane are transformed to spatial frequency components and the spatial frequency components and combination of this transform with quantization of the pixels eliminates information of high frequency component which is insensitively recognized by the human's eye and realizes high compression ratio of pixel data. In JPEG (Joint Photographic coding Expert Group) which is international standard for a method of still image compression encoding, DCT is used for transform encoding and in JPEG2000 wavelet transform is used for transform encoding.
A method, by which the original image data is encoded as mentioned above and then the predetermined data is written in a header and a control signal of a code, is also considered.
When the predetermined data is embedded in low-order bit of the gray scale value of the above mentioned image data, there is a problem of unavoidable deterioration of quality of the image in which the predetermined data is embedded.
In addition, when the predetermined data is written in the header and the control signal of the code, deterioration of the quality of the image can be avoided, but there is a problem that the predetermined data can be simply retrieved and an illegal copy of the original image data can be easily made because the predetermined data is written separately from the original image data.